Vocal acoustics may offer a clue in diagnosing ASD

Researchers in France, Italy, and Chile report that differences in vocal acoustics (for instance, frequency and pitch) may be helpful in identifying children with autism spectrum disorders (ASD). 

Frédéric Briend and colleagues compared 29 children with ASD to 20 typically developing (TD) children. In addition, they compared the group with ASD to 20 children with hearing loss who were fitted with cochlear implants, and to 15 children with developmental language disorder, because these conditions have commonalities with ASD. 

Asking the children to repeat non-word sounds, the researchers measured nine different acoustic properties in their speech samples. They then used machine learning to develop an algorithm to classify the samples. They report, “We showed that voice acoustics classified autism diagnosis with an overall accuracy of 91 percent against TD children, and of 85 percent against a heterogenous group of non-autistic children.” 

The researchers conclude, “Overall, our work suggests that easy-to-measure voice features, potentially linked to abnormal early neurodevelopment, can help in the diagnosis of autism spectrum disorder.” While their study involved children who were around eight years of age, they say future research should investigate the acoustic cry features of babies as a potential biomarker for ASD.

Citations

“Voice acoustics allow classifying autism spectrum disorder with high accuracy,” Frédéric Briend, Céline David, Silvia Silleresi, Joëlle Malvy, Sandrine Ferré, and Marianne Latinus, Translational Psychiatry, July 8, 2023 (free online). Address: Marianne Latinus, UMR 1253, iBrain, Université de Tours, INSERM, 37000 Tours, France, [email protected].